INDEX

Explanations

terms related to specific names or topics like "Kharma" and the name "Rahul" or "Manuel Molano."

oai_token-act-pair · gpt-3.5-turbo

references to specific characters or terms associated with a particular narrative or context

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

LC

-0.76

Rod

-0.67

Tit

-0.66

 Nost

-0.66

 Vest

-0.65

 Marketable

-0.64

 Mack

-0.63

HB

-0.63

Torrent

-0.63

 Stream

-0.63

POSITIVE LOGITS

arma

2.80

wal

1.52

PF

1.44

Rao

1.41

 Rahul

1.39

phal

1.36

 Sharma

1.34

qi

1.24

 Ripple

1.22

reek

1.20

Activations Density 0.052%

terms related to specific names or topics like "Kharma" and the name "Rahul" or "Manuel Molano."

references to specific characters or terms associated with a particular narrative or context

No Comments

No Known Activations